Skip to main content

Full text of "Physiography and Settlements in Southern Macedonia"

See other formats

Early Journal Content on JSTOR, Free to Anyone in the World 

This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in 
the world by JSTOR. 

Known as the Early Journal Content, this set of works include research articles, news, letters, and other 
writings published in more than 200 of the oldest leading academic journals. The works date from the 
mid-seventeenth to the early twentieth centuries. 

We encourage people to read and share the Early Journal Content openly and to tell others that this 
resource exists. People may post this content online or redistribute in any way for non-commercial 

Read more about Early Journal Content at 
journal-content . 

JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people 
discover, use, and build upon a wide range of content through a powerful research and teaching 
platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit 
organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please 


By Alan G. Ogilvie 

[With Fig. 4 on PI. II facing p. 178 and separate map, PI. Ill facing p. 196] 

In a previous paper 1 an attempt was made to elucidate in a general way 
the geographic basis of life in a section of Macedonia, and the discussion of 
the physical controls included a brief description of the land forms. Through- 
out the history of man in the Balkan Peninsula such a powerful influence 
has been exerted by the peculiar conditions of relief, and of soil and water 
distribution, that it seems to be worth while to emphasize at greater length 
in the present paper the character of these features and the human response 
to them; for, while the examination deals with but a small section of the 
Peninsula, the physiography of this region may be taken as typical of large 
tracts between the lower Danube and the Aegean. 

As explained in the former paper, the nature of my military duties in 
Macedonia precluded any thorough physiographic investigation; but I was 
called upon to make a number of journeys — mostly rapid — in many direc- 
tions throughout the area. Moreover, the detailed mapping of the country 
for military purposes 2 has permitted a more thorough study of the relief 
than has hitherto been possible. In so far, then, as the description of the 
landscape given below is explanatory, it is intended to be suggestive and in 
no sense final. My statements regarding the geology of the region are largely 
taken from the work of others (see appendix) — of the most of which I was 
ignorant when in the country; but I believe that my physiographic deduc- 
tions, based partly on the previous work but mainly upon my own obser- 
vations of the topographic features, may be of service to the geologist, who 
in the future will be able to treat of the subject after making a thorough 
geological survey. The present geographical paper only points to probable 
geological causes as a help to the description of the topographic features. 

The outstanding topographic features of the region are plateaus and 
flood plains. The plateaus exhibit a great variety of shapes in plan. Their 
surfaces are due partly to erosion and partly to deposition, and they are 
more or less dissected. Their edges are largely straight and are always 
abrupt. The flood plains which separate the plateaus are isolated save for 
narrow defiles which connect them. They are partly filled by lakes or by 
marshes — the relics of former lakes. 

J "A Contribution to the Geography of Macedonia," Geogr. Journ., Vol. 55, 1920, pp. 1-34- 
; The maps made during the war cover all the region north of the latitude of Saloniki. For most of Khal- 
kidhike I had no adequate map, and my plotting of features there cannot be regarded as accurate. 



The Rocks 

All of the high land and parts of the lower plateaus are underlain by rocks 
which are transitional in character between the purely crystalline of the 
Rhodope to the northeast and the sediments of the Dinaric ranges to the 
west. The most extensive outcrops are of green chlorite schists with veins 
of quartz, serpentine, etc., belts of gneiss and important bands of limestone, 
gray to white in color, and showing all degrees of alteration up to the white 
marble which composes most of the pyramid of Mt. Athos. In the eastern 
part of Khalkidhike there are two lens-shaped outcrops of a granite, so 
much disintegrated as to be relatively weak in resisting erosion. Along the 
major faults which run southeast from Saloniki — and probably on other 
faults also — are notable bosses of basic intrusive rocks, mostly gabbro and 
diorite. On the higher plateaus there are large stretches covered by a red 
residual clay, which bears witness to the long period of erosion to which these 
surfaces are due. 

The lowland corresponds mainly to the areas of clay left by the so-called 
Sarmatic sea — a huge fresh-water body which formerly covered the whole 
region — or of beds of gravel, sand, and fresh-water limestone deposited in 
the gradually shrinking lakes of the separate basins. The flood plains also 
contain important belts of alluvium, and, bordering the highest escarpments, 
there are large alluvial fans of well graded materials. 

Their Structure and Relation to Relief 

It will be seen below that difference of rock outcrop does not account for 
the largest features of relief. But in the uplands relative strength of the 
rocks has been an important factor in determining the highest residual ridges 
as also in locating the valleys more recently incised in the older surfaces. 
Generally the limestones and gneisses form ridges in all parts, and the pale 
limestone crags form a characteristic feature along a belt running from 
Dub (above Doiran) southeastward with numerous gaps through Deve 
Kran, Khortach Dagh, Kholomon Vuni, and northern Longos to Mt. Athos. 

In many parts of the Rhodope massif the metamorphic rocks exhibit a 
west-southwest-east-northeast strike, and in the eastern part of our region 
the rock outcrops frequently have this trend. Throughout the bulk of the 
area, however, the strike varies but little from the northwest-southeast 
direction. This is the trend of the Dinaric Alps to the west of the zone in 
question. The two strikes come into active conflict in the Krusha Balkan, 
eastern Beshik Dagh, eastern Khalkidhike, and the peninsulas of Athos and 
Longos. One result of the conflict may have been the foundering of the 
basins which lie on the margins of those areas. Otherwise the rock structure 
has had but little effect on the topography, save that differential erosion has 
etched the surface of the plateaus, as mentioned above, and that in con- 
sequence many of the minor watercourses and a few of the more important 
are "strike streams. " 


The most outstanding feature of the Macedonian landscape, the contrast 
between plateau and plain, has now to be discussed in the light of its origin. 
Briefly it is a legacy from the end of the Tertiary and beginning of the 
Quaternary periods when the whole of the central Balkan Peninsula was 
convulsed and rent by crustal disturbances which culminated in the depres- 
sion of the Aegean basin — previously a land area and probably similar to the 
remaining highlands— which formed the southern rim of the Sarmatic lake. 
By these convulsive movements the entire rock base of our region became 
partitioned into blocks, some of which sank while others rose or remained 
stationary. Of these blocks the former correspond to the basins, the latter 
to the plateaus. The separate basins continued to hold water, and as the 
level of this gradually sank the lakes became isolated. 

I have stated that the edges of the plateaus are abrupt and straight. They 
are indeed probably fault scarps in nearly every case, and, while they are 
deeply notched by streams, the ends of the spurs are in alignment. Their 
foot is frequently marked by a line of mineral springs of varying temperature, 
and in some cases bosses of igneous intrusive rock stand forward from the 
scarp like bastions. The scarps in the majority of cases truncate the various 
rock outcrops. In other words direction of scarp and direction of strike 
rarely coincide. It will be seen, then, that the greater features of relief, the 
block plateaus and the basins, are due not to normal processes of denudation 
but to the breaking up of the crust at a period so recent that subsequent 
processes have as yet altered the form given to the region in very small 

A close study of the relief reveals the fact that the outlines of the various 
plateaus and plains form a remarkable pattern in which there are many 
parallel sides, while the prolongation of a line bounding the foot of one 
escarpment is often found to lie along the foot of another escarpment. Many 
lines of this character have been drawn on the map — where they are de- 
scribed as "supposed fault lines." In few cases have I had the opportunity 
of proving their existence geologically. The lines have been drawn mainly 
from topographic considerations, and, where normal processes of erosion 
are regarded as insufficient to account for existing relief, the aid of recent 
faulting is invoked. In many cases this evidence is strengthened by the 
existence of warm springs on or near the lines as drawn or by reports of 
earthquakes in their vicinity. 

A more detailed discussion of the fault lines will be found in the appendix. 

Geographic Subdivisions 

The proper subdivision of the region for geographic discussion corresponds 
to the groups of block plateaus and basins; and in the description of land- 
scapes and human settlements given below this scheme is followed. 

The map (Fig. i) shows the outline of these main groups as well as that 
of interior topographic divisions to be referred to later. The highest blocks 



are indicated by a plus sign and the lowest by a minus sign. The arrows indi- 
cate the probable direction of initial tilt in some of the blocks. 

The first group of high plateaus, "A" (map, Fig. 1), occupies the northern 
and eastern limits of the region and contains in the Belashitsa Mountains 
the highest land (2,175 nieters) 
in the region. As we are con- 
cerned only with the edges of 
these plateaus they will be dealt 
with along with the trough which 
they overlook. 

The plateaus and basins of 
the Vardar, "B." This com- 
prises the basins of Gevgeli and 
the Saloniki Campagna sepa- 
rated by a rather complex sys- 
tem of small and mostly low 

The Krusha-Beshik plateaus, 
"C," with small included basins 
form the largest block mountain 
system in the region. It ex- 
tends from Lake Doiran south- 
eastwards across the mouth of 
the Struma. 

The Doiran-Struma trough, 
"D," lying between groups "A" 
and "C." 

The Khalkidhike group, "E," 
is made up of a complex system 
of alternating block mountains 
and basins with the prong of the 
Khortach Dagh extending north- 
west and the three fingers ex- 
tending southeast. 

The Langadha-Beshik trough, 3 
"F," lying between groups "C" 
and "E." Fig. 2 

+ A / 


ssrA. a + 

\ v. 

\. — 



Fig. i — Key map to crustal blocks and physiographic 

The Plateaus 

The plateaus then are upon fault blocks, and many of these are tilted, so 
that each has a steeper and a gentler slope. Their surfaces bear traces of 
long periods of erosion prior to their separation by faulting, and there is a 
remarkable accordance of level over wide tracts. The study of these 

• Accents, in this article, indicate the stressed syllable. 


relatively smooth surfaces is important both to the geologist who attempts 
to trace the physical history of the region and to the geographer, since the 
distribution of the population is largely determined by that of the available 
smooth surface. The most outstanding features of this kind are shown by 
tints on the map. 

Briefly there are four well marked surfaces, of which the two older — and 
the higher — lie on the upraised blocks and largely on crystalline rocks. 

The Lake Terraces 

The two younger surfaces may perhaps be best described as lake terraces, 
for they are found round the margins of the basins or on the slightly raised 
blocks which must have stood as shallows, with islands gradually emerging 
from the waters of the shrinking lakes. These lower surfaces are evidently 
due mainly to sedimentation and wave action. They consist largely of 
clays and sands, and the small amount of dissection which they have under- 
gone points to their recent origin. For convenience of reference I shall speak 
of them as the terraces, upper and lower. The upper terrace varies between 
280 and 240 meters in altitude. It has been identified at intervals all along 
the southwestern border of the Krusha Balkan block and on the northern 
side of Khalkidhike, as well as east of the Struma mouth and west of the 
lower Vardar. The lower terrace lies between the 180 and 140 meter 
contours and is found on both sides of the lower Vardar valley. 

The Upper and Lower Peneplains 

The two higher surfaces appear to be due mainly to the normal processes 
of erosion and aggradation. For convenience I shall refer to them as the 
upper and lower peneplains. Further research on the ground may assign 
them to a single period ; but there is in many places a marked break between 
the surfaces which may well indicate an interruption in the topographic 
cycle. Each of the peneplains appears to be highest in the northwest and 
lowest in the southeast, the higher ranging from about 780 to 480 meters, 
and the lower from about 420 to 320. 

The higher peneplain forms the summits of the spurs of the Belashitsa 
north of Lake Doiran and of the western part of the Krusha Balkan. It 
then widens southeastwards to the great saddle between the Krusha and 
Beshik blocks. This saddle of Lahana lies in a direct line between the gorge 
at the eastern end of the Belashitsa, where the Struma breaks through that 
range, and the pass "Dervend" in the Khortach block five miles north of 
Saloniki, which is undoubtedly the remains of a river valley. This fact led 
Cvijic with good reason to interpret both features as the work of an older 
Struma. But in each case the valley trend was probably determined by 
fault line weaknesses. The peneplain is also well developed south of Beshik 
Dagh and on all the higher blocks of Khalkidhike, where it has a very 
uniform altitude of about 500 meters. It is perhaps the most striking feature 


in the region although it is now dissected to a much greater extent than can 
be shown on a small-scale map. Its remnants are of the greatest importance 
to the population, as will be seen below. 

The remnants of the lower peneplain are most continuous on the south- 
western side of the Krusha and Beshik blocks, but it has also been identified 
on the block mountains of Khalkidhike. It ranges from about 420 to 320 
meters in altitude, possibly lower in southeastern Khalkidhike. It is more 
dissected than the higher peneplain except where it reaches back from the 
margins of the blocks, as at the headwaters of the Galiko and in the enclosed 
basin of Mavrovo where it is a surface of aggradation. 

Above the highest peneplain there rise the crests of the block mountains. 
While the rocks forming these represent the roots of some older mountain 
system, they are now simply residual ridges on the higher peneplain. They 
fall into four groups: the Belashitsa Planina; Krusha Balkan and Beshik 
Dagh; the Khortach, Kholomon, and Stratonikon ridges in Khalkidhike; 
and the crests of the two fingers of Longos and Athos. 

The parts of the map which remain untinted are either the wide alluvial 
flats of the basins, often ill-drained and containing the various lakes and 
marshes; or else they represent slopes between the plateaus or regions too 
much dissected for plateaus to be recognizable. I did not visit the south- 
western pait of Khalkidhike and the two western fingers, so these areas are 
left untinted. 

The Doiran-Struma Trough 

This important depression falls naturally into western and eastern 
sections. The western section is narrow — from three to ten kilometers 
wide — and is walled on the north by the abrupt and almost straight fault 
scarp of the Belashitsa range and on the south by the deeply dissected slopes 
of the Krusha Balkan, only less steep than the northern wall (see Figs. 3 
and 4.) It is a funnel very difficult of access from either wall. Its western 
extremity is occupied by Lake Doiran, almost circular and flat- bottomed, 
beyond which it is all but hemmed in by the low plateaus of the Vardar. 
While the bed of the funnel bears traces in its alluvial deposits of a former 
river draining a greater Doiran to the Struma, the watershed is now situated 
ten kilometers east of the lake and some 100 meters above it. This is the 
col of Dova Tepe, above which on a spur to the south stands the modern 
fort of that name. 

The Belashitsa front exhibits three distinct surfaces. The highest part is 
smooth, very steep, and bare of trees. It represents a residual ridge on the 
highest peneplain. At from 300 to 400 meters below the crest project a 
series of gently sloping spurs offering summer pastures. These are relics 
of the highest peneplain. They are separated by deep gullies covered with 
deciduous scrub or forest. The spurs themselves end in abrupt bare facets — 
parts of the fault scarp proper. The main gullies are fronted by large 
alluvial fans, which continue to be built up while the smaller streams from 

£ x °. 

<fi h j3 

Q jy -c 

» « « 

O "S ■£ 

■5 o c 

•fi £ 2 

52 « 

cu c c 

is O I 

m . o 

S m bo 

•5 3 fi 

<L> K/ "^ 

* co T3 

QJ — 2 

bfl J3 -fi 

g .a T3 

.5 .fi c 

rf £ <tf 

- ° 8 

! ^4 C 

O 2 ° 

fi © 2 -S 5 3 

O T3 W 

o rt^ 

2 Q o 

^ O to 

o x *■; 

5 x - 

bo I 

JS ^ S 

a> cu S 
•5 •* ea 

a o H - 

cu •- -is 

a fi -o 

— cu fi 

CO £ 

fi <u o 

fc c3 •£ 

•fi co cy 

*i a) »h 

» g s 

fi°l s 

.2 « <fi 

; « w 5 ^ 
: « •§ S 1 1 

it- 8 1 1 "8 

fi <u ti "5 rt 

S. g fi o 2 

1 J^ •-{ 4-. CO ^ 

1 cj ra >_ -j-> -. 

rt "O i3 

"2 * fi w 

> fi ^ 

v Q ^ 

P< >» 2 -S : 




.13 o 

•* S». ^? 

^ O fi 

S ° 

« % I 

pa 3 2 

J TJ .2 

* fi T3 

« rt <U 


5 H 

fi *_> 


o o 

" 4J o fi 

O 3 <V <u 

2 ^ £ 

£ £ o 

c <y 
7 "5 ^ 

5 ^ 

J -fi 
1—1 hit 

v« CU « 

<u fi .2 

w fi ^ "fci 

♦j H c o 

. bfl o 

s •§ & 


Fig. 4 — Panorama from a northwestern spur of the Krusha Balkan taken 

Top panel: In the background, the mountains of Payik (left) and Meglen (right), lying between the Vardar and Cherna valleys. It was along this ri 
ridge, seen in elevation. The Bulgarian army occupied the summit of this ridge for over two years. 

Center panel: Portion of higher peneplain with main Bulgarian defense system on the fault .scarp (left); western end of Belashitsa Planina (right); 
Bottom panel: Part of Belashitsa Planina with fault scarp marked by facetted spurs; irrigated alluvial fans in foreground. 

The Geographical Review, Vol. XI, No. 2, 1921, PL II 

ur of the Krusha Balkan taken about midwinter. For extent see Fig. 2. 

ia valleys. It was along this ridge that the Serbian army finally broke through. In the foreground, the lake and town of Doiran backed by the Dub 

of Belashitsa Planina (right); zone of villages across the trough (left foreground). 


the spurs have incised their beds in the fans. The rivers draining the trough 
to the east and the west hug the southern rim, forced thither by the fans. 

The village sites are of two types characteristically placed in relation to 
the alluvial fans. 4 The first group, consisting of large villages, lies close up 
to the fault scarp, each village at the apex of a fan, commanding the water 
which is diverted into hundreds of irrigation channels radiating on the fan. 
The other group occupies a zone stretching across the trough at the lower 
edge of the fans where the ground water emerges, but above the swampy 
zone around Lake Doiran. These villages are small, but there are eleven of 
them in the zone between the trough walls, which are here six kilometers 
apart. The small town of Doiran stands majestically on the southern shore 
of the lake, built partly on a narrow terrace and partly on the steep slope of 
the Dub ridge. The site, which controls traffic on the old road from the 
Aegean to the upper Vardar, has been occupied since very early times. The 
town, which normally has over 3,000 inhabitants, has always provided the 
bulk of the lake fishermen. 

The eastern portion of the trough differs from the western mainly in its 
much greater length — about 80 kilometers — and bread th— about 12 kilo- 
meters on the average — and in the flatness of its floor. Its northern wall, the 
Belashitsa, has characteristics similar to those of its western section. The 
northeastern wall consists of high mountain blocks cut by deep gorges; but 
throughout most of the length a much-dissected lake terrace, widest in the 
east, intervenes between wall and plain. This terrace is also found in the 
south, but not along the western wall, which is abrupt and cut by numerous 
canyons, many of which are impassable. 

The only routes by which the basin is easily accessible are the western 
funnel, by which the railway enters it; the saddle (east of the region) by 
which it leaves, leading to Drama and Constantinople; the deep defile by 
which the Struma enters from the north and the much shallower defile by 
which it seeks the sea; the Lahana saddle, central on the southwestern 
wall; and a higher saddle central on the northeastern wall. Each of these 
is followed by a road. 

The whole basin is rimmed by alluvial fans; and, as the more important 
streams come from north and east, the Struma is forced towards the south- 
western wall. This river itself fills the northeastern corner with a large, flat 
fan which dams the waters of the western funnel, forming the lake and 
swamps of Butkova. 

This lower Struma plain is one of the most important agricultural areas in 
Macedonia, and most of its surface is under permanent cultivation. The 
population — mainly Slav in the north and Greek in the south — occupies 
villages whose size corresponds generally to the fertility of the soil. Again, 
we find one group at the tops of the fans — the town of Seres being by far the 
biggest of these, another set of large villages farther out and drawing their 
water mostly from shallow wells, and a third group of small places close to 

4 The map showing settlements (Fig. 10) may be consulted. The map is discussed on pp. 193-195. 


the marshes. A notable feature of the landscape of the plain is the clumps 
of magnificent elm trees grouped about the villages, which give much needed 
shade in summer and provide nesting places for the storks. 

The southern extremity contains the last remnant of a great lake. But 
the shallow Ahinos is fast filling up and could now be artificially drained 
with little difficulty. The small loss to the fisherman would be amply repaid 
by the gain to agriculture and the removal of the curse of malaria. 

The study of population density is instructive in this plain where every 
foot of cultivable land is utilized; and it is safe to assume that the statistics 
of 1905 — which happen to be available — show approximately the population 
which such land can support with the local methods of agriculture. In the 
map illustrating the distribution of settlements (Fig. 10) I have outlined the 
plain and divided it into two zones, the first (A) including the fans and drier 
belt, the second (B) covering the wetter ground but omitting the lake and 
swamps. The population per square kilometer in zone A is 217 and in 
zone B 66, or 561 and 170 per square mile respectively. In the computation 
one-half of the population of Seres, or 4,650, was disregarded, as being not 
directly dependent on the land of the plain. 

The Krusha-Beshik Blocks 

This highland area extends for 100 kilometers from northwest to southeast 
with an average breadth of nearly 30 kilometers. The highest parts of the 
Krusha Balkan lie on two blocks set nearly at right angles to each other 
(Fig. 1). These overlook — on the south and west — a large triangular block 
with its apex to the south and corresponding to the drainage basin of the 
Galiko River (Fig. 3). Before the scarp of the northern high block there are 
the two smaller triangular basins of Snevche and Rayanovo. Of the two 
high blocks the western has a narrow crest with a fairly uniform altitude of 
about 700 meters, while the eastern contains a wide stretch of the highest 
peneplain and has residual ridges rising to 1,175 meters. These high crests 
are covered with deciduous forest. The Galiko and its tributaries, while 
they are separated by a number of steep and rocky spurs, flow for the most 
part in valleys incised in the lower peneplain. There are therefore extensive 
smooth surfaces, with much residual clay, almost entirely cleared of forest 
and given up to intermittent agriculture. But the steep valley sides are 
covered with brush and are difficult to cross. The basins of Snevche and 
Rayanovo are filled with a thick red clay deposit which the streams have 
somewhat dissected. Here the agriculture is permanent. 

The valleys of the Galiko and its main tributary, the Spanc, contain a 
few relatively important villages. Otherwise the most striking feature in 
the distribution of settlements is the bunching of small villages in a few 
areas, which are frequently so broken by gullies that the extent of cultivable 
soil cannot have been the attraction. The reason for the close grouping of 
the villages may lie in the need for their original inhabitants to combine 



for defense. This at least would seem to be the case in the groups east of 
the Galiko, which are Turkish : every hamlet is so placed that all windows 
can command the approaches, which are usually steep and narrow. 

The most astonishing village swarm occurs on the northern slope of the 
Krusha Balkan, where there are 48 situated in an area of about 125 square 
kilometers, or between two and three per square kilometer. Of course 
these villages are small, but the slope is steep — it falls 600 meters in some 
three and a half kilometers. The explanation probably is that the natives 
obtain a livelihood by farming in the trough below; but, as the villages 
were evacuated during the war, I could not find out if this was actually 
so. At the southwestern edge of the lower block and overlooking the Vardar 
basin is a wide patch of the higher lake terrace with edges falling steeply 


Fig. 5— Lake Beshik area. On the left, the plateau of Ogurli (O) and terraced valley of Koja Dere with 
Pazarkia (P) on the delta; on the right, summits of Beshik Dagh behind, with Suho (S), Mavrovo enclosed 
basin with lakes in front. All the settlements are shown. 

away on three sides. At its center is a butte — a miniature relic of the 
lower peneplain. On the butte is a church, and at the foot of the butte 
lies the town of Kukush (Kilkish) formerly with some 10,000 inhabitants. 
It derives its importance from its nodal position and is a collecting center 
for the produce of the plateaus to the north of it. 

Between the Krusha and Beshik Mountains lies a continuous stretch of 
the higher peneplain broken by the deep ravines of the rivers which run 
northeast and southwest from the saddle of Lahana. Except for the steep 
ravines, which are scrub-covered, all of this plateau is marked by inter- 
mittent cultivation. Its chief geographical importance is as a route, and 
it must have served in this way for many centuries. One of the few pre- 
historic sites away from the basins is found near the village of Berovo and 
14 kilometers from the plain, while the remains of an important 
classical Greek town lie at the junction of the route with the Struma valley. 
The existing villages are remarkable for their size. They are wide apart 
and consequently the farmers travel far to their fields. 

The two masses of the Beshik Dagh contain the widest areas of land above 
the high peneplain in the region. These summits are entirely covered with 


deciduous forest or brush. The scarps facing south and west are precipitous, 
while the northern slopes are very deeply dissected and have an extremely 
small population. Between the high scarps and the Beshik Lake the surfaces 
of the two peneplains are wide and little dissected, and they have mostly 
been cultivated at one time or another (Fig. 5). The only access to these 
plateaus is by difficult trails up the fault scarp from the south and through 
the gap north of Mavrovo. The small basin in the eastern part was formerly 
drained to the southeast, but the two small lakes Mavrovo and Lanja now 
have no outlet. The plateau is dominated by Suho, a large, closely built 
village of over 4,000 inhabitants which stands at the upper edge close under 
the high scarp; the other villages are small. 

The rugged block of the Beshik Dagh has two extensions, one eastward 
joining it to the Bunar mountains, beyond the region; the other, southward, 
linking it to the Khalkidhike. The first of these thresholds is severed by the 
Struma mouth. The river seems to have cut a wide valley across it at the 
level of the lower peneplain and then, since the subsidence of the Aegean 
basin, to have cut its existing narrow passage nearly to the sea level. The 
hill within the incised meander of the defile is the site of the ancient Greek 
city Amphipolis, the "City [with the river] about" [it], and before its foun- 
dation the site is mentioned (by Herodotus) as the "Nine Ways" — clearly 
a nodal point of importance. There are three passages through the south- 
ward extension of the Beshik block. Two of them are at high levels, 340 and 
140 meters respectively, and may mark previous drainage outlets. The 
large village of Vrasta marks the point of their junction on the seaward 
side. The deep, wooded defile of Rendina carries the present drainage 
channel from Lake Beshik with a fall of only 20 meters from lake to sea. 

The seaward scarps of the Beshik block mountains are abrupt and are 
scarcely dissected by the steep mountain torrents which seam them. The 
higher parts bear deciduous forest and bush while the lower slopes are clad 
in a luxuriant mantle of the Mediterranean macchia, which is here at its 
northern limit. There is a wide beach which carries fine groves of olive trees 
north of Vrasta and of planes about Stavros at the mouth of tHe Rendina.. 

The Langadha-Beshik Trough 

From its western threshold at Deve Kran this trough extends for 30 
kilometers to the southeast with a nearly constant width of 10 kilometers. 
It then forks, the southeasterly direction being continued for another 20 
kilometers at the level of the higher lake terrace, while the northern and 
deeper fork turns east-northeast for 10 kilometers with a width of 5 kilo- 
meters and then east for 23 kilometers more with a width of 10 kilometers. 
The trough is bounded almost everywhere by abrupt parallel fault scarps 
which lead directly to portions of the higher or lower peneplains. The forking 
above described is due to the isolation of a roughly triangular block, which 
I have called the Ogurli plateau from a village on its surface (Fig. 5). The 


highest part of the block exhibits a portion of the higher peneplain still un- 
dissected and containing a small lake. This level area forms the farm land of 
two villages. The slopes are covered with bush. The lake terrace which 
extends from side to side in the southern fork of the Langadha trough is 
nearly flat and carries a shallow lake which has no outlet but is in imminent 
danger of being drained either eastward or westward by the head streams 
which are cutting back from both sides. The terrace is cultivated inter- 
mittently over most of its surface, but there are extensive patches of high 
thorny Paliurus bush. 

The same fault which isolates the Ogurli block seems to be responsible 
for two other features of importance. It has brought up the hard limestone 
which forms the abrupt toothlike crag of Deve Kran (570 meters) stand- 
ing guardian of the western threshold of the trough, and it gives outlet 
to a constant supply of hot mineral waters east of Langadha village, where 
throughout history the baths have attracted the Salonikans. This fault 
and those related to it formed the focus of important earthquakes as late 
as 1902. 

The natural scrub vegetation of the trough walls is very sparse in the west 
but gets progressively more luxuriant eastward — an evidence of increasing 
humidity in that direction. All the bottom land, however, which is not 
swamp is cultivated and mainly without irrigation. The two residual lakes 
are very shallow except for the eastern part of Beshik, which is 22 meters 
deep. They contain considerable quantities of fish, and fishing provides 
occupation for several villages. The only outlet from Langadha is an 
artificial canal. Except in the western end of the trough and in the flat delta 
which extends northward into Beshik, the villages are situated on the 
terraces or fans round the trough walls; and on the eastern side of the 
Ogurli block there is a line of six places built on consecutive patches of 
dissected lake terrace. The southern edge of the trough is the reputed route 
of the Via Egnatia, the Roman highway across the Balkan Peninsula; while 
the six prehistoric and two classical Greek sites hitherto identified bear 
witness to the long period of human occupation. 


The high kernel of this peninsula consists of the long narrow blocks of the 
Khortach Dagh, extending from the Galiko River for 60 kilometers to the 
southwest, and the Kholomon Vuni, running northeast from the end of the 
Khortach Dagh for 20 kilometers and then after a gap continuing eastward 
as the Stratonikon Oros to the sea. 

All of these have much land above the higher peneplain, and the two last 
named carry the most important deciduous forests in the region. To the 
southwest and south of these are several blocks at a lower elevation and 
tilted in various directions (see map, Fig. 1), while the three blocks forming 
the fingers can be considered as separate units. The Khortach blocks will 

1 84 


be discussed first, then western and eastern Khalkidhike in order, and 
finally the three fingers. 

Khortach Dagh 

The parallel sides of the block are fault scarps, and the block as a whole 
is a horst (Figs. 6 and 8). Cross faults at right angles to the others account 
first for the broken northwestern end of the block — a scarp of which the 
precipitous character is emphasized by the Galiko defile at its foot ; secondly, 
for the sudden rise in elevation above the high peneplain seven kilometers 
east of Saloniki; and thirdly, in all probability for the hollow just north of 
the city and for the old river valley known as the Dervend or "pass" now 


Fig. 6 — Khortach Dagh near the western end. On the left is the summit of the ridge (1208 m.), Khor- 
tachkeui (Kh.) on its plateau, and Lake Langadha with the fishing village of Aivasil (A.). Kirechkeui (K.) is in 
the valley behind the limestone monadnock (center). On the right is the gulf and city of Saloniki (S.) and the 
Dervend pass in the foreground, with the mouths of the Galiko on the extreme right. Nearly all villages are 

followed by the only highway across the range and probably also the course 
of the Roman Via Egnatia. The surface of the block for the first 30 kilo- 
meters of its length forms part of the lower and the higher peneplains ; and 
in each case the hard limestone bands form the residual ridges, and the 
surface is deeply incised by ravines. The surface is mostly rocky and arid, 
and the villages dotted around the flanks are mainly concerned in cultivating 
the ground lower down. Kirechkeui, the "limestone village," with nearly 
9,000 inhabitants, divides its attentions between the quarries above it and 
the cultivation of the slopes of the sheltered "strike" valley in which it lies. 
Khortachkeui lies on a patch of the high peneplain, flat as a table from edge 
to edge of the block and overlooked by the scarp of the forested ridge 
(1,208 meters) southeast of it. The inhabitants farm the residual clay of 
the plateau while from the forests above they furnish Saloniki with firewood 
and charcoal as well as chestnuts in the autumn, and, in the summer, with 
ice cut from artificial reservoirs on the mountain. Springs in the scarp also 
provide one of the main sources of water supply for Saloniki, and the 
plateau carries a magnificent aqueduct dating at least from Byzantine times. 




The position at the head of a deep gulf of the Mediterranean coinciding 
with the entrance to the main north-south corridor through the Balkan 
Peninsula has always been occupied by a relatively large settlement, and 
the important prehistoric site lying to the southeast of the city bears witness 
to the earliest occupation. The nodality of the position was enhanced by 


Fig. 7 — Saloniki and environs. From a map by the French Service Topographique de FArmee; i, citadel; 
2, Turkish town; 3, Arch of Galerius; 4, White Tower; 5, prehistoric settlement; 6, Kalamaria; 7, Kapujilar; 
8, Monastir and Athens railway; 9, Nish and Belgrade railway; 10. Constantinople railway; n t road to Ki- 
rechkeui; 12, road to Seres (via the Derbend); 13, Vardar mouths. Contour interval 50 meters. 

the construction of the Via Egnatia, which first touches the Aegean at this 
point. In the choice of the special site, the point of contact between the 
Khortach ridge and the sea, the aim must have been to control the traffic 
alongshore. The remains of the prehistoric settlement lie on the dissected 
shore terrace and are bounded by a ravine (map, Fig. 7) . But a hill fort must 
have been built at an early date on the last spur of the Khortach block 
200 meters above the sea, and the city must have grown downwards to meet 
a maritime settlement below. By the time the hill fort had been converted 
into the Crusaders' citadel (which now serves as a civil prison), and the 


containing walls had been built down to the sea, the city must have had its 
present size excluding the Kalamaria suburb; indeed it may have had this 
size even in classical times. This epoch has given to Saloniki the rectangular 
plan of the lower town with its main artery, the Egnatia Street, which is 
spanned at its eastern end by the triumphal arch erected about 300 A.D. in 
memory of Galerius, a general of Diocletian. This form is retained in the 
closely packed Jewish quarter. Whatever the original plan of the steep 
Turkish quarter below the citadel, it is now a typical medieval Turkish town 
with narrow irregular streets and stairs but with all the charm of the 
picturesque combination of overhanging houses, shady interiors, and quiet 
leafy courtyards and with delicate minarets always in the view. The 
devastating fire of August, 1917, left this part untouched but cleared the 
entire southern quarter and greatly reduced the accommodation of the city 
already overcrowded by its 170,000 inhabitants. The site of Saloniki, while 
fine from the point of view of medieval defense, is such that modern exten- 
sions have had to be northwestward for commercial needs along the radiating 
routes, and southeastward in the residential suburb of Kalamaria. Beyond 
the suburb the land rises, and a bluff thrust forward towards the fingered 
delta of the Vardar is crowned by a modern guardian fort. Inland on the 
same rise lies the old Turkish village Kapijilar, which signifies "the keepers 
of the gate." 

The high part of Khortach Dagh has its smooth fault scarp on the north- 
west and its ledge of the higher peneplain broken only by narrow ravines; 
but in the southwestern face the smoothness is broken by two bastions 
standing forward — the bosses of igneous rock lying on the fault line. The 
few villages on the slopes of Khortach are mainly pastoral but with patches 
of cultivated land on the peneplain remnants. In general, however, the 
higher slopes are covered with deciduous bush or forest while lower down is 
the humble "pseudomacchia" of dark-leaved evergreen oak. 

At the foot of this high block lies the low triangular plain of Vasilika, open 
to the gulf and narrowing eastwards (Fig. 8). This rich bottom land is all 
under permanent cultivation, and it supports Vasilika, a place of 1,500 in- 
habitants, and a line of smaller villages situated along the foot of the south- 
ern wall. If another reminder of the import of geological faults be needed 
here, it is found in the hot mineral baths at Sedes close to the remains of two 
large prehistoric settlements. East of Vasilika is a rocky threshold leading to 
the almost inclosed basin dominated by the town of Galatista with about 
2,000 inhabitants. The basin is limited on three sides by high, straight, and 
very steep scarps clad with scrub. The bottom of this basin seems to form 
part of the higher lake terrace, while along it to north and south are frag- 
ments of the lower peneplain. All the gentle slopes are cultivated, at least 
intermittently. Galatista is built partly on the steep rocky slope and partly 
on a terrace of fresh-water limestone. The steep slope below is terraced, 
irrigated, and intensively cultivated, while the flat land of the basin is 
farmed without irrigation. The southern scarp of the basin has also a band 


I8 7 

of terraced fields with many fruit trees, but these are owned and tended by 
the inhabitants of Vavdhos, a village situated some 200 meters above on the 
high plateau — another instance of the uneconomical mode of concentrated 
village life in Macedonia. 

Western Khalkidhike also includes the high block of Vavdhos, almost a 
rectangle, with steep scarps facing the Galatista basin (west) the Resitnik 
basin (north) and tilted eastward to form a lower threshold joining it to the 
Kholomondos block. The western end of the latter carries the southeast- 
ward trend of the features down to the base of the Longos peninsula with a 
high steep scarp facing southwest in front of which lies Poliyiros on the 
higher peneplain. The Vavdhos threshold is deeply trenched by the 

THE GEOGR. REVIEW, April 1921 

Fig. 8 — Western Khalkidhike. On the left is the eastern end of Khortach Dagh with the Galatista (G) basin 
leading to the Vasilika (V) plain. In the center are the massifs of Kholomon Vuni (behind) with Poliyiros (P), 
and of Vavdhos, with village of the same name (Vs). These are divided by a gorge leading from the basin of the 
Resitnik (R). Three igneous bosses are situated left of Vasilika, right of Vavdhos, and right of Poliyiros. On 
the extreme right is the isthmus of Cassandra, the site of Potidaea. All important villages are shown. 

incised meanders of the Resitnik Dere, which thus drains the oval basin of 
that name southwards by an impassable defile. This is probably an example 
of an antecedent river, and another case equally striking is found in the 
Miyalka, the principal river of eastern Khalkidhike cutting through a 
similar threshold by a similar defile north of the Longos peninsula (Fig. 9). 
The fault which limits these blocks on the southwest is marked by two fur- 
ther bosses of igneous rock each forming a group of hills, the one close up to 
the Vavdhos massif and the other near the northern shore of the Cassandra 

The whole of the remaining land consists in a block gently tilted south- 
wards from the scarp overlooking the Vasilika-Galatista basins. This 
block consists in the north of crystalline rocks with a north-northwesterly 
strike, but over most of the area lies a thick mantle of weak sediments — 
clays for the most part, trenched by a series of valleys widely open near the 
sea and narrower near the crest. The coast is mostly low and arid, and there 
are three wave-built cusps, two of them near the western extremity, and the 
third forming the isthmus which links the finger of Cassandra to the main- 
land. Whatever may have been the original appearance of this sloping 
plateau of the southwest, it now looks like a rolling steppe; for, while it has 


all been cultivated, there is always so much lying fallow that only the valley 
bottoms have the appearance of farm land. 

The bulk of the Khalkidhike population is on this plateau, distributed in 
some 35 villages of which the smaller are occupied by Mussulmans, probably 
Yuruks from Anatolia, the larger population being Greek. There are 
also a number of metochia, or farms, belonging to the Athos monasteries and 
worked by the monks. The steep valley sides and the higher parts of the 
plateau are covered with evergreen-oak scrub. The Resitnik Dere widens 
near its mouth to a flood plain which extends eastwards along the Gulf. 
Here the natural vegetation is true macchia bush, but much of it has given 
place to fine groves of olive and mulberry, to vineyards, and other typically 
Mediterranean cultures. About the valley of the Resitnik the population 
is segregated in a small number of widely separated settlements: the high 
village of Vavdhos with its pastures and some intermittent cultivation about 
it and its lower farms in the Galatista basin; Resitnik in its upper basin, 
with access to Galitista through a pass between the Khortach and Vavdhos 
massifs; Poliyiros, the old capital of Khalkidhike, central as to position but 
difficult of access; Ormilia at the head of the flood plain at the mouth of 
the Miyalka; and two villages on the flood plain of the Resitnik. Of these 
Poliyiros is the biggest (c. 2,000). It was founded probably in classical 
times on account of ancient gold-mining attraction but now exists by 
agriculture both on its own plateau, where it has the highest olive groves in 
the region, and near the sea where it possesses a group of huts for the 
temporary residence of its farmers. Ormilia (Fig. 9) has its agricultural 
plain and a fishing station on the Gulf two and a half kilometers away, while 
its inhabitants also provide miners to work the chromite and magnesite 
which occur about the igneous boss aforementioned, the minerals being 
taken directly to sea. 

At intervals along the shore of western Khalkidhike the mounds of 
prehistoric settlements are a feature of the landscape. The ancient Greek 
town of Potidaea was situated at the isthmus of Cassandra, and Olynthos 
and Mekyberna lay just east of the Resitnik Dere mouth. Their importance 
may perhaps be judged from the fact that Potidaea appears to have stretched 
from sea to sea, and Olynthos to have occupied nearly three square kilo- 
meters of area. 

Eastern Khalkidhike 

This hilly country is, humanly speaking, one of the most remote in 
Europe (Fig. 9) ; for, now that its gold which seems to have attracted the 
ancients to it no longer repays the working, its ways are little trodden by 
outsiders. It leads nowhere save to Athos, and it is much easier to reach 
that by sea. But its remoteness makes it worthy of study. Its build accounts 
for its inaccessibility. The crustal blocks which compose it are so oriented 
and tilted that there is a central basin; and, although the lowest wall of this 
is on the southeast, the coast there is a recent fault scarp, and the streams 


drain inland almost from the precipitous coast itself. The drainage outlet 
from the central basin is the deep and winding antecedent defile cut by the 
Miyalka through upturned strata of hard limestone; and this is virtually 
impassable. The eastern crustal block consists of schists in the north and 
altered granite in the south; it is tilted towards the Gulf of Erissos, and 
its seaward slope is deeply dissected by streams, only one of which reaches 
the sea, the others burying themselves in the sands of the flat coastal plain. 
The block forming the Stratonikon Oros in the northeast has an abrupt 
straight scarp facing south and extending right out to Cape Eleftera and 
seems to be tilted northward to meet the extension of the Beshik block in an 
east-west fracture line. This area is extremely broken by erosion ; but the 


Fig. 9— Eastern Khalkidhike. In the background are the ridges of Kholomon (left) and Stratonikon (right) 
with the higher peneplain extending from Larigovo (L) to Izvoron (I). The central basin is drained by the 
Miyalka through its gorge to the plain of Ormilia (O). In front of a fault scarp are the isolated plateaus of 
Revenik (R) and Gomati (G). On the right is Erissos (E) and the beginning of the Athos peninsula with the 
remains of the Xerxes canal (X) on the isthmus. 

line of contact is probably indicated by the through valley drained east and 
west, which is marked by a continuous strip of the lower peneplain surface. 

The higher ridges of eastern Khalkidhike are covered with forests — 
probably primeval — of oak, beech, and chestnut, and the lower hills are clad 
in bush derived from similar forests which have been cut and which have 
been prevented from recovery by browsing goats. The valley bottoms are 
lined with fine groves of plane trees, and the coastal strips bear a luxuriant 
growth of macchia made up of numerous species of varied type. On the 
southern coast the Mediterranean pine forest (Pinus halepensis) has main- 
tained itself in patches lending a pleasing contrast to the landscape. Land 
form and vegetation combine to make eastern Khalkidhike one of the 
beauty spots of Macedonia. 

About the high ridges considerable tracts of the two peneplains remain 
intact, and it is on these that the main population is found. The strip of 
the higher surface which follows the southeastern slope of Kholomondos 
and that which faces the Gulf of Erissos are too deeply dissected to be 
inhabited; but to the east of the Kholomon range there lies a smooth 


plateau surface with a heavy residual soil and an area of at least 20 square 
kilometers, not yet dissected by the erosion which is attacking the edges 
from three sides. This plateau is farmed or grazed — for there is much of it 
under grass — from three villages, the chief of which is Larigovo. An east- 
ward extension of the plateau runs nearly to the sea but here is cut into 
patches, on one of which stands Izvoro close under the scarp and command- 
ing a magnificent view of the pyramid of Athos in the southeast. It is a 
village of miners, and a light railway carries the ore — lead, antimony, and 
gold-bearing pyrites — to the gulf near by for export. 

Under the fault scarp of the eastern block lie two patches of level land, 
portions of the two peneplains. The higher, which is dissected by narrow 
ravines leading to the central basin, is the domain of Revenik and is all farm 
land ; the lower flat seems to be the upper part of an old wide valley, perhaps 
even a lake bed, of which the main part has disappeared with the subsided 
crust; while the stream by its violent erosion headward from the new gulf 
has all but broken upon the placid beauty of this little plain of Gomati, with 
its trim fields and sluggish plane-girt stream, and the village nestles in the 
farthest recess of the plain as if in fear of being undermined. 

The central basin contains many hilly outcrops of the crystalline rocks, 
but its tributaries provide much flat land. The basin seems to be strikingly 
underpopulated, and for no apparent reason, for it is healthful, sheltered, and 
well watered. It is divided from the sea by a wide strip of bush-covered 
plateau, part of the lower peneplain, from which rises a higher wooded 

The isthmus of Xerxes, which joins the Athos Peninsula to the mainland, 
rests upon the sunken end of the crystalline block northwest of it, and the 
rocks of this appear on the western coast; but the isthmus consists mostly 
of weak sediments dipping gently to the east. On a northern spur of a table 
of these stands Erissos, the successor of the classical Akanthos, and there is 
evidence that this strong strategic position has been occupied continuously 
since very early times. The coastal plain north of the village is fertile and 
moist, and in it all kinds of Mediterranean crops are raised; while the 
undulating land on the isthmus gives admirable pasture. Moreover, 
Erissos has its skala, or landing place, with many fishing boats. The remains 
of the canal cut by Xerxes 2,403 years ago to permit the passage of the 
Persian armada have been surveyed by Struck, 5 and the canal was found by 
him to have a length of some 2,450 meters, while its highest point is some 
14 meters above the sea. 

The Three Fingers 

Of Cassandra, Longos, and Athos 1 have visited only the last. The others 
I have seen only from afar. But it seems clear that all three owe their 
remarkable shape to the subsidence of crustal blocks about them. Cassandra 
appears to consist mainly of weak sedimentary rocks, while Longos, like 

5 Adolf Struck: Makedonische Fahrten, I, Chalkidike, Vienna and Leipzig, "1907. 


Athos, is almost entirely crystalline in character. Both are cut by many 
deep ravines and are crowned with Mediterranean pine forests, while the 
lower slopes have a macchia cover. In each there are considerable tracts of 
farm land, probably most in Cassandra. The villages and metochia are 
numerous and small, but it is doubtful if the population of Cassandra — which 
is Greek, like the most of Khalkidhike — is even now as great as it was in 182 1, 
when 700 families with their flocks and herds are reported to have been 
massacred by the Turks after an insurrection. The natural vegetation of 
both peninsulas is very favorable to bee keeping, an occupation much 

The Holy Mountain (Agion Oros) or peninsula of Athos presents one of the 
most striking landscapes in the Mediterranean. The long, narrow, rugged 
ridge, rising gradually towards its end and then springing suddenly at the 
tip to the pyramidal summit of white marble 1 ,935 meters above the sea, 
would be picturesque if it were all bare rock, but it is clad in a dense mantle 
of vegetation save for the high peak itself; luxuriant macchia bush and pine 
forests below, oak and chestnut higher up, and finally dark fir near the 
pyramid. Moreover, all along both sides twenty great monasteries and 
twelve subordinate houses occupy the most surprising and often most 
inaccessible sites, some of them perched on rugged spurs, others nestling in 
wooded chasms, but all of them beautiful in their varied medieval archi- 
tecture. The great Russian monastery of Panteleimon with its white walls 
and green copper domes might be a piece of the Kremlin transported to a 
very different but still suitable setting; the Greek house of St. Paul at the 
western foot of Athos itself with the whiteness of its Byzantine walls vying 
with that of a huge fan of marble blocks which a recent earthquake detached 
from the pyramid above and hurled down the ravine to clog its mouth below 
the monastery — these are but two of the many astonishing glimpses to be 
had from the western gulf. On the eastern side, which as a rule is somewhat 
less abrupt, many of the monastic centers have had more room to spread 
their buildings in a picturesque irregularity; such places are Vatopedi, 
which is Greek, and the monastic village of Karies, the seat of the Holy 
Synod — the government of the monkish republic — while Khilandar, the 
Serbian house in the northeast, unlike most others lies out of sight of the 
sea in a secluded valley; but its defense was ensured by a high keep near the 
shore, whose summit is visible from the monastery. 

The Vardar Basin 

The lower Vardar valley which forms the western margin of the region is 
made up primarily of two lowlands separated by a higher block which is 
traversed by the Vardar in the antecedent Tsinganska Klisura, "Gipsy 

I have had opportunity only for the most cursory observations in this 
section, and the following outstanding features of its geography have been 


gleaned mainly from the writings of Cvijic, who has examined it in detail, 6 
and from study of the new maps. 

The high crustal blocks in the north overlook several lower blocks, nearly 
level plateaus, which are attributable to the various surfaces of erosion 
referred to throughout. They are composed in the west of granites and 
elsewhere of folded crystalline rocks, whose strike is northeast-southwest 
save in the high Dub ridge of Doiran, where it is north-northwest-south- 
southeast. The strength of the rocks here accounts for the isolation of the 
ridge. These plateaus have been attacked by erosion along lines of weakness, 
but the striking feature is still the extent of level surface broken only by 
narrow ravines. This smoothness is emphasized by the absence of any 
vegetation save the humblest oak scrub and the grass which becomes burnt 
up in the long summer drought. 

The basin in the extreme northwest of the region appears to be due to 
crustal subsidence, like the Doiran depression, from which it is completely 
separated by the higher blocks; but the basin of Gevgeli, with which it is 
connected by the Vardar valley, is mainly due to normal erosion. The 
characteristic soil of all these plateaus and basins is generally sandy and 
coarse. In part it is residual and in part deposited, but it is always relatively 
infertile; and the large agricultural settlements are found near the Vardar 
with their fields on the alluvial fans brought from the well-watered moun- 
tains to the west or on the residual clays of the hill slopes. There is no lack 
of water for irrigation in the basin, and the large villages are surrounded by 
vineyards and mulberry groves, for it is one of the leading European centers 
of silk production. 

West of the Vardar defile on patches of the peneplain, here trenched by 
deep ravines, where their inaccessibility renders them strong in defense, 
stand two of the largest Turkish villages in the region — each with over 2 ,000 
inhabitants. Otherwise the plateau villages are small. 

South of the Vardar defile lie the great flood plain, known as the Cam- 
pagna of Saloniki, and the marshes of the delta. The plain has an even 
fault scarp for its western wall, but on the east it has two long extensions. 
The first of these runs to the northeast, is bounded by straight scarps of the 
crystalline rock on all sides, and all but reaches Lake Doiran. The low 
threshold separating the two basins has been cut through by the headward 
erosion of a Vardar tributary; with the result that the lake, which at a 
higher level was drained to the Struma, empties into the basin now under 
discussion, except in summer when there is no outflow. The bottom of the 
basin is occupied by Lake Arjan, by its marshes and peat beds, and by the 
deposits of a former greater lake which are now somewhat dissected by 
streams. The population lives either on the margins of this basin, where 
water emerges from the rocky rim, or else in small villages close to the 
streams. Elsewhere lack of permanent water precludes settlement, but the 
whole surface is cultivated intermittently. 

6 Jovan Cviji6: Grundlinien der Geographie und Geologie von Mazedonien und Altserbien, Petermanns Mitt. 
Erganzungsheft No. 162, 1908. 


South of this basin stands a wide plateau, about 100 square kilometers in 
area, of folded crystalline schist and limestone, ill-drained and thinly 
populated save at one fertile spot. The plateau is mostly given up to 
grazing. It overlooks in turn the second great arm of the Vardar plain, 
which leads southeastward through an area of alternating aridity and 
undrained swamp — the bottom of the former lake — to the Galiko valley. 
This section, like the last, has a very small population. But south of it 
stands an oval plateau of sedimentary strata round the margin of which is 
a ring of villages and a similar ring of prehistoric sites. The population 
probably lives by raising poor crops from an arid soil and by grazing flocks. 

The Vardar has built an immense elongated fan beginning at the Gipsy 
Defile and reaching nearly 20 kilometers southwards. This accounts for the 
damming back of Lake Arjan. But the river has cut through its own fan 
and now runs in a very variable channel, frequently overflowing its banks. 
As a result the villagers on the east of the river valley do not practice 
intensive agriculture to any extent but are largely occupied with cattle 
keeping, for the pastures of the flood lands are rich. Moreover, in addition 
to the permanent settlements, a feature of the landscape is the temporary 
huts built of reeds from the swamps and occupied by the Vlach shepherds, 
who migrate hither from the hill pastures in the winter. Similar conditions 
regulate the life on the Vardar delta, where there are numerous villages and 
many temporary settlements amid the maze of reed beds. 

The western rim of the Campagna presents an entirely different aspect. 
Here the villages are large and stand well above the plain on the terraces 
and well- watered alluvial fans. It is a zone of rich permanent agriculture 
with an important silk and wine production. The settlements of the 
Campagna thus fall into three classes: a western strip of important per- 
manent agriculture by irrigation; a valley and delta strip with flooded 
pastures and cattle; and an eastern area with widely scattered villages 
practicing intermittent agriculture and tending flocks over a wide range of 
poor pasture. 

Population Distribution 

Figure 10 is a map showing the settlements of the region. The population 
statistics have been taken mainly from a source 7 which cannot be considered 
unbiased as regards race but is probably as good as can be found respecting 
the numbers. These statistics disregard the Mussulman population, which 
I have had to infer from various sources. My aim has been to show the 
population as distributed prior to the first Balkan war, in 191 2, the wars 
which have involved the region almost continuously since that date having 
considerably upset the status and in particular caused a considerable evacu- 
ation of Turks from the region. In the Great War the population near the 
front was of course evacuated, and many villages were destroyed. In order 
to illustrate the meaning of this I have inserted the line which was the 
"front" in the region from i9i6to 1918. 

7 D. M. Brancoff : La Macedoine et sa population chretienne, Paris, 1905. 



In Macedonia people live entirely in villages, chiefly from motives of 
security in a troubled land. A map which shows every village therefore 
gives an accurate representation of the distribution of population. 

I have added to the map the sites of known prehistoric settlements. 8 The 
remains of these are interpreted as the habitations of a people who lived to a 
large extent upon shellfish and were therefore dependent on the sea. They 


Fig. io 

would be further restricted to the coast lands by the forests which, es an- 
alysis of the vegetation shows, must have been much more extensive, at least 
in the eastern part of the region ; and these peoples could not clear forests. 
The chief classical and Byzantine towns which are known to history and 
archeology have also been shown. Their distribution is probably determined 

s Taken from Adolf Struck: Makedonische Fahrten, I, Chalkidike, Vienna and Leipzig, 1907; Bull, de 
Correspondance Hellinique, Vol. 9, 1916, Paris; The Annual of the British School at Athens, No. 23, Session 1918- 
1919, London. 


by considerations similar to those which have affected the modern popu- 
lation, but doubtless the dependence upon sea communication with Greece 
accounts for the predominance of large maritime settlements. 

The most striking feature of present-day distribution is that the main 
mass of the people live or depend on the basins or the plateaus composed of 
sedimentary rocks. These are the areas where agriculture is simplest. The 
next is the surprisingly large population which exists by scratching the thin 
soil over every possible space on the high dissected plateaus. 

The population will always remain overwhelmingly agricultural and 
pastoral; but, given a settled and progressive government, there must 
assuredly be certain definite changes in the population. The chief of these 
are as follows. The swamps and perhaps part of the lakes will be drained, 
additional rich farm land being thus afforded and malaria banished. Irri- 
gation will be considerably extended, and dry- farming methods will be 
introduced in the more arid parts, while antiquated implements will be 
replaced by modern machinery. The hill populations will then have 
sufficient incentive to move down to the plains or to the lower plateaus, 
which with present methods are too arid to support many people. With 
scientific agriculture the Struma plain, for instance, would probably support 
nearly 500 people per square kilometer, instead of 217 as it does in its best 
section today. The hills will be reserved for pasture and forests; but of 
course these two interests must be kept quite separate, for the flocks and 
above all the goat must be kept out of the forest reservations if these are to 


In drawing the fracture lines shown on the map attention has been paid 
first to the geological boundaries, secondly to such information as exists 
regarding earthquakes in the region, thirdly to the distribution of hot and 
mineral springs, and lastly to the study of the topography. The information 
regarding the first three is contained mainly in the works of Neumayr, 
Burgerstein, Philippson, Cvijic, Oestreich, Hoernes, and Nopcsa. 9 

9 M. Neumayr: Geologische Untersuchungen iiber den nordlichen und ostlichen Theil der Halbinsel Chal- 
kidike, Denkschriften Kaiserl. Akad. der Wiss. in Wien, Mathem.-N aturw. Classe, Vol. 40, 1880, pp. 328-339. 

Leo Burgerstein: Geologische Untersuchungen im siidwestlichen Theile der Halbinsel Chalkidike, ibid.. 
Pp. 321-327. 

Alfred Philippson: La tectonique de l'Egeide, Ann. de Geogr., Vol. 7, 1898, pp. 112-141. 

Jovan Cvijic: Die tektonischen Vorgange in der Rhodopemasse, Sitzungsber. Kaiserl. Akad. der Wiss. in 
Wien, Mathem.-N aturw. Classe, Vol. 90, 1901, pp. 409-432. 

Idem: Grundlinien der Geographie und Geologie von Mazedonien und Altserbien, Petermanns Mitt. 
Erganzungsheft No. 162, 1908. 

Idem: L'ancien lac egeen, Ann. de Geogr. Vol. 20, 191 1, pp. 233-259. 

Idem: Die Tektonik der Balkanhalbinsel, Compte Rendu Congres Giol. Internatl. IX, Vienna, 1903, pp. 

Karl Oestreich: Beitrage zur Geomorphologie von Makedonien, Abhandlungen K.K. Geogr. Gesell. in Wien, 
Vol. 4, 1902, pp. 1-169. 

Idem: Die Oberflache Mazedoniens, Geogr. Zeitschr., Vol. 16, 1910, pp. 560-572. 

Rudolf Hoernes: Das Erdbeben von Saloniki am 5 Juli 1902 und der Zusammenhang der makedonischen 
Beben mit den tektonischen Vorgangen in der Rhodopemasse, Mitt, der Erdbeben-Kommission der Kaiserl. 
Akad. der Wiss. in Wien, No. 13 (N.S.), 1902. 

F. Baron Nopcsa: Die Mineralquellen Makedoniens, Mitt. K.K. Geogr, Gesell. in Wien, Vol. 51, 1908 
pp. 242-292. 


As regards topographic evidence I have assumed that where a straight 
escarpment exists which cannot be explained as the ordinary boundary 
between weak and resistant rocks, a fault runs along somewhere near its 
foot. In most cases the slope is taken to be a fault scarp — the dissected 
fault face — and not the fault line scarp due to differential erosion of unequal 
rocks thrown together by faulting. In this respect there is strong similarity 
of appearance to the scarps of the Great Basin of the North American 
Cordillera. In a few cases a fault is presumed to exist where one or more 
river valleys have a straight course not parallel to the rock strike, but 
these lines have only been drawn where there was different evidence for 
another section of the same line. 

The fractures are marked on the map as continuous lines where there is 
evidence of one kind or another, and in some cases two such lines are joined 
by a broken line to indicate the supposed continuation. The probable 
direction of "throw" is indicated where possible by the usual spurs on the 
line, and it is noticeable that the throw along any fracture is now to one 
side, now to the other. In several cases two parallel faults have been drawn 
where only one may turn out to exist. 

The fracture lines may be grouped as follows: 

1. The "A" group, with an average northwest-southeast trend, i.e. 
parallel to the Dinaric folding. 

2. The "B" group, with an average northeast-southwest trend, i.e. 
parallel to the pre-Dinaric folding. 

3. The "C" group, of which the trend varies little from east- west. 

4. The "D" group, of which the direction varies little from north-south. 
Several curved faults are shown, and it is possible that other lines which 

meet at a small angle may prove to be curves. Following is a list of the 
fractures numbered as on the map. The evidence for the existence of each is 
stated thus: Geological, "geol.;" hot springs, "h.s.;" straight scarps, "sc.;" 
straight river valleys, "val. :" 

"A" group, beginning in the west: 

A 1, sc, and h.s. in N 

A 2, sc. and geol. in N and S sections; val. in center 

A3, sc. and geol. 

A 4, geol. and sc. in N; h.s. near Langadha; sc. Sof Ogurli; may be continuous with 
sc. on S coast 

A 5, sc. and h.s. in N, possibly geol. at Deve Kran 

A 6, sc. Truncates Olympus range at SW edge of region 

A 7, sc. and geol. (igneous intrusion) and h.s., perhaps continuous to Ormilia (geol.) 

A 8, sc. and geol. (intrusions) 

A 9, sc. 

A 10, continuous sc. 

A 11, sc. at Langadha perhaps continued to Revenik sc. and beyond 

A 12, sc. and h.s. 

A 13, sc. and h.s. 

A 14, (in Beshik block) sc. 

A 15, (in Beshik block) sc. 

A 16, (in Beshik block) sc. and val. 


Higher terrace, 240-280 
Lower peneplain, 320 -420 
Higher peneplain, 480 -720 
High residual ridges, 720-2175 
Large alluvial fans 





Copyright J92L by the American Geographical Society ofNe* York 
The Geographical Review } VoL U , No 2 ,1921, Plate M 


"B" group, beginning in the northwest: 

B 1, B 2, B 3, B 4, and B 5, sc. 

B 6, sc. and h.s. 

B 7, sc. in SW; on Khortach and Krusha also val. 

B 8, val. These two converge on Struma defile which is on one of the main N-S 
fractures of the Balkan Peninsula 

B 9, geol. N of Saloniki, val. in Dervend and Lahana saddle 

B 10, sc. 

B 11, val. in SW; curving in NE sc. and h.s. 

B 12, sc; with step faults sc. and val. 

B 13 and B 14, a small trough 

B 15, sc. 

B 16, sc. connecting with curve to S, sc. 

B 17, B 18, B 19, B 20, and B 21, sc. 

"C" group, beginning in the north: 

C 1, C 2, and C 3, nearly continuous, sc. 

C 4, four separate sections, sc. 

C 5 and C 6, nearly continuous sc. 

C 7, sc. in W, val. in E 

C 8 and C 9, sc. 

C 10, sc. and geol. 

C 11, a curved fracture, sc. 

C 12, sc. 

"D" group, beginning in the west: 
D 1, sc, perhaps a curve with A 
D 2, sc. and val. 
D 3, sc 

D 4; sc. at Mavrovo and Ogurli, probably connected with line of Resitnik Dere, val. 
D 5, sc. at Mavrovo 
D 6, sc. and val. ; perhaps continuation determined situation of Miyalka defile